2058238 - 12.12 - 2.27% tresize + 2 more (OSX, Windows) regression on Tue July 21 2026

Status: RESOLVED FIXED

(Firefox :: Address Bar, defect, P2)

Description

[Performance Sheriff Bot]

Perfherder has detected a talos performance regression from push 6ce73127512654d7c24036bd8a39782e56eb88d2. As author of one of the patches included in that push, we need your help to address this regression.

Please acknowledge, and begin investigating this alert within 3 business days, or the patch(es) may be backed out in accordance with our regression policy.

Regression	Test	Platform	Options	Absolute values [old vs new]
12%	tresize (doc)	macosx1470-64-shippable	e10s fission stylo webrender	7.30 ms -> 8.18 ms
2%	tp5n nonmain_normal_fileio (doc)	windows11-64-25h2-shippable	e10s fission stylo webrender-sw	564,302,216.54 -> 578,146,431.00
2%	tp5n nonmain_normal_fileio (doc)	windows11-64-25h2-shippable	e10s fission stylo webrender	568,957,113.08 -> 581,868,324.00

Improvement	Test	Platform	Options	Absolute values [old vs new]
3%	tresize (doc)	linux1804-64-shippable-qr	e10s fission nova stylo webrender	21.96 ms -> 21.26 ms
3%	tresize (doc)	linux1804-64-shippable-qr	e10s fission stylo webrender	21.65 ms -> 21.08 ms

Support & Actionable Steps

For questions, please contact [e-mail]. You can also find help on Slack (#perf-help) or Matrix (#perftest).

Action guide details:

• See the guide to handling regression bugs for investigation steps.
• Review the alert summary for affected tests, graphs, and comparisons.
• Check the PerfCompare results.
• Tests on Try: Run all tests using ./mach try perf --alert 51507 (See documentation link for details).
• Test locally: To run the tests locally use ./mach perftest 51507
• Profiling jobs: Trigger jobs from Treeherder job view or request them from [e-mail].

Comment 1

[:Vincent]

Just a note that we're looking into this, but I'm still somewhat new to Firefox dev, and this is my first performance regression, so I'm still getting my bearings :)

Comment 2

[:Vincent]

Attachment: Bug 2058238 - Use a media query as a proxy for URL bar width. r=#urlbar-reviewers

r=#urlbar-reviewers

While technically, it's the size of the URL bar that matters, checking
that with a container query caused a performance regression (in the
tresize test on MacOS). Thus, we use a media query as a good-enough
proxy for URL bar width.

This PerfCompare run shows that it resolves the ~12% regression on
MacOS:
https://perf.compare/compare-lando-results?landoInstance=lando-prod-2025&baseLando=73863&newLando=73864&baseRepo=try&newRepo=try&framework=1

Comment 3

[Kaitlyn]

Out of curiosity, when a new feature introduces a measurable performance regression (like the container query here), who ultimately decides whether the regression is acceptable versus requiring an alternative implementation? Is there an established process for making that tradeoff?

Comment 4

[Mark Banner (:standard8)]

(In reply to Kaitlyn from comment #3)

Out of curiosity, when a new feature introduces a measurable performance regression (like the container query here), who ultimately decides whether the regression is acceptable versus requiring an alternative implementation? Is there an established process for making that tradeoff?

https://www.mozilla.org/about/governance/policies/regressions/ is the overview, though doesn't really make it clear (I believe it is generally on a per-team basis, but I don't know if the "owners" of the particular metrics should be involved). Probably worth asking in one of the perf-help channels.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/cc6d9ebf3740
https://hg.mozilla.org/integration/autoland/rev/adf2310f5eb5
Use a media query as a proxy for URL bar width. r=urlbar-reviewers,desktop-theme-reviewers,dao

Comment 6

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/adf2310f5eb5